Glovegate: how lab gloves can inflate microplastic measurements

A recent study by chemist Madeline E. Clough and colleagues at the University of Michigan has triggered a lively discussion in the microplastics research community. The researchers showed that some particles previously interpreted as microplastics may actually have come from a very ordinary laboratory item: disposable gloves. More precisely, the strongest false signals were produced not by the glove polymer itself, but by stearates, compounds used during glove manufacturing. The study, published in Analytical Methods, found that dry contact with common nitrile and latex gloves could produce an average of around 2,000 false microplastic signals per mm², while cleanroom nitrile gloves reduced this to roughly 100 false positives per mm².

So how did it happen that one of the contamination sources was literally in scientists' hands, and does this mean the microplastics problem has been exaggerated?

Measuring the almost immeasurable

The story began when young PhD researcher Madeline Clough was analysing samples of urban air from Michigan. She found microplastic concentrations around 1,000 times higher than values reported in the literature. She immediately suspected that something was wrong with the result and repeated the experiment, but the numbers remained just as high.

In this type of experiment, air is pumped through a filter, the collected material is transferred onto a substrate, and individual particles are examined under a microscope. Their chemical composition is then identified using spectroscopic methods. To distinguish microplastics from ordinary dust, pollen, soot, and other particles, researchers usually rely on Fourier-transform infrared spectroscopy (FTIR) or Raman spectroscopy.

The first result is often expressed as the number of microplastic-like particles per unit area of the substrate, for example, particles per square millimetre. That is still an abstract laboratory value. Researchers can then combine it with the size of the sample and convert it into a volumetric concentration of microplastics in air or water. That second value has a clearer physical meaning and can be compared with other studies.

But can it really be compared so easily?

First, the spectrometer usually examines only about 1–2% of the total substrate area. The rest must be extrapolated statistically. Even if particles are distributed evenly, and they often are not, this introduces uncertainty.

Second, particle size matters. The upper boundary of what counts as microplastic is generally agreed: 5 millimetres. The lower boundary is much more difficult. Instruments may fail to detect the smallest particles, often around 5 micrometres or below, or may count several tiny particles as one larger particle.

Third, the detection method itself has limitations. Earlier comparative studies showed that FTIR and Raman spectroscopy can produce different particle counts, especially for smaller particles. FTIR may miss more of the smallest particles, while Raman can be more sensitive but also more affected by fluorescence and noisy spectra. In some samples, methodological differences can shift results by orders of magnitude.

And this is before adding environmental variability. In animals and humans, microplastic concentrations in biological samples may depend on recent exposure, diet, and sample type. In water bodies, results depend on depth, currents, and sampling location. In the atmosphere, weather matters: wind can resuspend more particles into the air, while rain can wash them out.

Why do scientists monitor microplastics?

The concern is their potential impact on human health and ecosystems.

Many common plastic polymers are relatively chemically inert, which is why plastics are used in medical equipment, packaging, and household products. But microplastics are not just "tiny pieces of inert material." Their size allows them to reach places larger fragments cannot. They can also contain additives, weathering products, or adsorbed pollutants.

In 2019, the World Health Organization concluded that there was not enough evidence to show a clear human health risk from microplastics in drinking water, but also stressed that the evidence base was limited and more research was needed.

Since then, the field has moved quickly. A 2024 study in The New England Journal of Medicine found that patients with micro- and nanoplastics detected in carotid artery plaques had a higher risk of a combined endpoint of myocardial infarction, stroke, or death during follow-up. This does not prove causation, but it strengthened concerns about possible links between plastic particles, inflammation, and cardiovascular disease.

Other studies are exploring reproductive and intergenerational effects. For example, recent work on mice suggested that paternal exposure to microplastics can alter sperm small non-coding RNA profiles and affect offspring metabolic health.

Microplastics can also act as carriers for other contaminants. Their surfaces can adsorb toxic substances and potentially modify how those substances move through ecosystems.

So even before "glovegate," comparing microplastic studies was difficult. In early 2026, another global analysis suggested that atmospheric microplastic concentrations and emissions may previously have been overestimated by large margins, with modelled atmospheric concentrations 100 to 10,000 times lower than some earlier estimates. The same study found that land sources emit more than 20 times more microplastic particles to the atmosphere than the ocean, although ocean-emitted particles may contribute more mass because they are larger.

Hands on the table

When Clough saw the anomalously high concentration, she was probably not surprised that something was wrong. But she and her supervisor, Anne J. McNeil, were not willing to accept complete chaos in the results. They began changing one experimental parameter at a time while keeping all others fixed.

Soon they identified another source of false signals: laboratory gloves.

The main contaminant was not the glove polymer itself. It was stearates, compounds used during glove production. FTIR and Raman spectrometers could mistake them for microplastics, especially for polyethylene.

How could this happen?

Chemically, "microplastic" is not a single substance. It is a category that includes many polymers, including:

• Polyethylene, (C₂H₄)ₙ
• Polypropylene, (C₃H₆)ₙ
• Polyethylene terephthalate, (C₁₀H₈O₄)ₙ
• Polyvinyl chloride, (C₂H₃Cl)ₙ
• Polystyrene, (C₈H₈)ₙ
• Synthetic rubber

Each polymer produces a different infrared or Raman spectrum. Research papers usually report both the total number of detected particles and the share of each polymer type.

Stearates are salts or esters of stearic acid, with the general structure C₁₇H₃₅COOR. Unlike most plastics, many stearates are water-soluble or soap-like and do not behave environmentally like persistent plastic particles. Sodium stearate, for example, is one of the main components of soap.

The problem is the long hydrocarbon "tail" of the stearate molecule: C₁₇H₃₅–. To FTIR and Raman instruments, this tail can look very similar to parts of polyethylene and polypropylene molecules because both methods respond strongly to C–H bonds.

Stearates are commonly used in glove manufacturing. They help the finished glove separate from the mould. As a result, they are mostly found on the inner surface of the glove, but when a person puts the glove on, the residues can transfer to the outer surface, the side that touches samples, tools, filters, or substrates.

To test this, Clough and colleagues compared seven types of gloves from different manufacturers: three nitrile gloves, three latex gloves, and one cleanroom glove type. Cleanroom gloves are used in fields such as microelectronics and are specifically washed to reduce residues such as stearates.

The researchers rubbed each glove type against aluminium substrates with a force of 30 newtons, then analysed the substrates using microscopy, FTIR, and Raman spectroscopy.

Under the microscope, the substrates contained many particles that looked similar in shape and size to ordinary microplastic particles. The key difference appeared only during chemical identification.

When the instrument searched against a database that included stearates, the particles were correctly identified as stearates. But when the instrument searched only against a microplastics database, which is likely how many routine analyses are done, a large share of the stearate particles were classified as microplastics.

The issue lies partly in the quality of spectra and partly in the recognition algorithms. In a standard workflow, the spectrum of an unknown particle is compared with a reference library, and the best match is selected. If the match score exceeds a chosen threshold, the result is accepted.

But microplastic particles are small, irregular, weathered, dyed, or contaminated. Their spectra often have a low signal-to-noise ratio. Because of this, researchers may use relatively modest match thresholds, sometimes around 70%. Under those conditions, stearates can easily "disguise" themselves as microplastics.

The experiment showed that standard gloves added an average of around 2,000 false microplastic-like signals per mm² for both FTIR and Raman methods, with FTIR slightly higher in some cases. Cleanroom gloves performed much better, with fewer than 100 false signals per mm² on average. In the worst case, one glove type generated around 7,000 false signals per mm².

These numbers should not be compared directly with volumetric microplastic concentrations in air or water. On their own, particles per square millimetre of substrate are not an environmental concentration. What matters is the trend: depending on glove material and brand, false-positive results can differ by almost two orders of magnitude.

One latex glove brand released far fewer stearate particles, almost as few as the cleanroom glove. The reason remains unclear. The manufacturer confirmed that stearates were used in production and that no special cleaning procedure was applied.

This is especially interesting because many microplastic review papers had recommended using gloves to avoid contamination. Latex and nitrile gloves were often specifically recommended. A German research group had already pointed to stearate contamination in 2020, but that warning did not receive broad attention. Earlier work focused mainly on wet transfer, while Clough and colleagues showed that dry contact alone is enough.

Living with uncertainty

To make microplastic analysis more accurate, Clough and colleagues propose a more careful approach to spectral classification, including a method they describe as conformal prediction.

Instead of forcing the instrument to produce one "correct" answer, this approach estimates uncertainty in the match between the unknown spectrum and possible reference materials. It can better distinguish stearates from real microplastics, especially when both are present. The trade-off is time: the method is slower, and some particles must be classified as "uncertain."

The authors also note that FTIR can perform better than Raman for this specific problem because FTIR spectra more clearly show the carboxyl group that distinguishes stearates from polyethylene-like polymers. But FTIR still struggles more with the smallest particles. Researchers therefore face a methodological trade-off rather than a perfect solution.

Neither conformal prediction nor FTIR can fully eliminate uncertainty. The best practical option is to minimise contamination at the source: avoid glove contact with microplastic samples whenever possible, use tools rather than fingers, and switch to properly tested cleanroom gloves when gloves are necessary.

That is exactly what Madeline Clough plans to do when she returns to analysing Michigan air samples.

The University of Michigan study does not show that microplastics are absent from the environment. On the contrary, microplastics were still found even in samples that were not deliberately contaminated. Nor did the study suddenly reveal that all earlier microplastic research is useless. Scientists and environmental chemists have been aware for years that microplastic measurements are difficult to compare across methods, particle sizes, sampling protocols, and environmental conditions.

What the new work does is add one more important source of uncertainty: the gloves.

For a field that already struggles with sampling bias, small-particle detection, polymer identification, weather variability, and inconsistent reporting standards, this is not a minor detail. It is a reminder that microplastic science needs not only better instruments, but better contamination control, better reference libraries, clearer uncertainty reporting, and more reproducible workflows.

Microplastics are still real. The problem is still real. But the numbers may be less straightforward than they look.

This article is also published on N+1, in Russian. illuminem Voices is a democratic space presenting the opinions of leading Sustainability Thought Leaders, their views do not necessarily represent those of illuminem.

The world needs sustainability knowledge. At illuminem, no interest group or shareholder can influence our work. Thank you for supporting our mission to make high-quality and independent sustainability information free for all. Every contribution helps. Thank you for donating today.